📜 ⬆️ ⬇️

Comparison of WIPmania and MaxMind GeoIP databases

In connection with the alternative that appeared to the base, MaxMind decided to compare how much these bases differ. It is the differences that are compared, not the correctness / incorrectness of the definition.

Comparison

1) The size of the bases. MaxMind base is about twice as large as WIPmania (100k vs 50k)
2) The number of ip-addresses entering into all ranges is practically the same here and there.
3) Address ranges are 88% identical. (That is, start / end ip-addresses in the database of MaxMind and WIPmania)
4) Of these 88%, only 662 (1.2%) of the subnets reassigned the country.
5) Next, the number of ip-addresses entering into all subnets for the country.
Country, maxmind IP, wipmania IP,% difference

  + ------ + ------------- + -------------- + -------- +
 |  cc |  maxmind_ips |  wipmania_ips |  diff |
 + ------ + ------------- + -------------- + -------- +
 |  AZ |  157186 |  228845 |  0.31% |
 |  BY |  334595 |  335737 |  0.00% |
 |  EE |  1015116 |  1018665 |  0.00% |
 |  KZ |  615802 |  612209 |  -0.01% |
 |  LT |  1943069 |  1970561 |  0.01% |
 |  LV |  1349597 |  1390552 |  0.03% |
 |  RU |  19126288 |  20374227 |  0.06% |
 |  UA |  3748921 |  4148605 |  0.10% |
 + ------ + ------------- + -------------- + -------- + 


Tests

Compare how much different the WipMania data from MaxMind is for uneven ranges.
Algorithm such:
a) we define the country by the initial and final ip in the WipMania database by the MaxMind database.
b) we get data like start_ip, end_ip, wipmania_cc, maxmind_start_cc, maxmind_end_cc
c) select all subnets where maxmind_start_cc! = wipmania_cc OR maxmind_end_cc! = wipmania_cc
The result is 2633 entries.
e) besides, they were undecided on the basis of GeoIP 1394
The difference is about 7%.
')
Findings:
1) Indeed, other techniques have been used to determine GeoIP.
2) The relevance and accuracy of this data, like the data of any other geoip service, is difficult to establish.
3) Good luck to the creators of the project, I am glad that a good alternative to MaxMind has appeared.

Source: https://habr.com/ru/post/38711/


All Articles